JPH06324896A - Method and device for correcting error - Google Patents

Abstract

PURPOSE:To utilize the capability of a duplex coded error correction code to a maximum. CONSTITUTION:Error detection/correction is performed by C1-decoding the second code block of a duplex coded code word by a second inspection word, and a flag is added to the second code block in which an error is detected. Thence, the error detection is performed by releasing interleave and by C2- decoding a first code block by a first inspection word with minimum distance d1, and when the number of erroneous words shows a number <=S shat is 2S1<=d1-1, the erroneous word is corrected. When the number of erroneous words exceeds S1, and when relation between the number E2 of known erroneous words provided with the flag and the number S2 of unknown erroneous words without possessing the flag is shown as 2S2+E2<=d1-1, and the number up to S2 of erroneous words other than the known E2 words is detected by C2- decoding again the first code block assuming that E2 errors exist, and the number up to S2+E2 of errors can be corrected.

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to error correction in a digital signal reproducing apparatus, and particularly to an error suitable when a double-coded error correction code is used. A correction method and apparatus. [0002] When transmitting a digital signal, a check word is added at the time of transmission in order to cope with a data error occurring in a transmission system, and error detection and correction are performed by the check word at the time of reception. I often do it. Reed-Solomon code, which is efficient and easy to decode, is often used as the check word. In particular, the method of performing double encoding by the Reed-Solomon code can obtain excellent correction capability and is used in digital audio disks and the like. FIG. 1 shows an example of double encoding by the Reed-Solomon code, where 1 is an information word, 2 is a first check word, and 3 is a second check word. In the encoding, first, the first check words Q ₀ , Q ₁ , Q ₂ and Q ₃ are added to the 12 information words W _{0 to} W ₁₁ indicated by the diagonal broken lines in FIG. The block consisting of 12 information words and 4 first check words is C
₂ blocks 5 Next, the symbols of the C ₂ block 5 are interleaved to form an array as shown in FIG. Then, a data block consisting of twelve information words W _{0 to} W ₁₁ selected from different C ₂ blocks and the first check words Q ₀ , Q ₁ , Q ₂ and Q ₃ (in FIG. The second check words P ₀ and P ₁ are added to the (viewed data block). A block composed of these 12 information words, 4 first check words and 2 second check words (that is, a data block for one horizontal row in FIG. 1) is referred to as a C ₁ block. When decrypting,
Deinterleaving is performed after error detection and error correction is performed on the C ₁ block, and error detection and error correction is performed on the C ₂ block. In the C ₁ block, the code length is 18, the number of check words is 2 and the minimum distance is 3, and in the C ₂ block, the code length is 1
6, Reed-Solomon code with the number of inspection words 4 and the minimum distance 5 is used. Therefore, in C ₁ decoding, error correction of 1 symbol is possible (that is, the number of check words 2 indicates the location of error data (by one check word) and the content of the error (by one check word). ). Also, C
_{In 2} decoding, error correction is performed within a range of 2S + E ≦ 4 for S errors whose error positions are unknown and E errors whose error positions are known (in the following description, the former is an error and the latter is an erasure). Can be carried out (Japanese Patent Application No. 58-110931)
issue). Therefore, in C ₁ decoding, error detection and 1 symbol correction are performed, and at the same time, a flag indicating the decoding state is added to the C ₁ block, and in C ₂ decoding, the flag added in C ₁ decoding. By performing the optimum decoding of the following three kinds of decoding according to the situation of (3), it is possible to perform error correction with excellent capability. (1) S = 2 E = 0: Two errors are corrected. (2) S = 1 E = 2: Correct one error and two erasures. (3) S = 0 E = 4: Correct 4 erasures. FIG. 2 is a flowchart showing C ₁ decoding. In the figure, N (E) is the number of errors detected by C ₁ decoding, and when it is judged that there is one error, one symbol correction is performed. Also, two flags (F ₀ flag as a flag indicating the state of decoding of a C ₁ decoding,
F ₁ flag). The F ₀ flag is set to "1" when an error is detected, and the F ₁ flag is set to "1" when two or more errors are made and the correction is impossible. FIG. 3 is a flowchart showing C ₂ decoding. In the figure, N (F ₀ ) and N (F ₁ ) are
These are the numbers of F ₀ flags and F ₁ flags set in the C ₂ block as seen in the direction of the oblique broken line in FIG. 1, respectively. In C ₂ decoding, the error situation is estimated by the number of flags and the optimum decoding method is determined. For example, when N (F ₀ ) = 4, that is, when the number of F ₀ flags (including those with the F ₁ flag added) is 4, the position where the flag is added is incorrect. As the position, the erasure correction of (3) is performed. FIG. 4 shows an example in which there are three errors, where 6 is a C ₁ block and 7 is a C ₂ block.
Also. 8 indicates an erroneous symbol to which a flag determined for each C ₁ block is added, and 9 indicates a correct symbol to which a flag determined for each C ₁ block is added. In the figure, in C ₁ decoding, if there is an error, a flag is added to all the symbols in the block, but if the error is 1 symbol, it will be corrected correctly, and if there are 2 or more errors. But there are correct and incorrect symbols. Therefore, there are an erroneous symbol with a flag added and a correct symbol with a flag added. [0015] By the way, the above (3)
In erasure correction of, although up to 4 errors with flags added can be corrected, the error detection capability is limited,
If there is an error to which a flag is not added, erroneous correction may be performed. FIG. 5 shows an example thereof. 10 is an erroneous symbol to which no flag is added. The erroneous symbol 10 to which this flag is not added is due to missed detection in C ₁ decoding. Accordingly, as shown in FIG. 1, C ₁
If the number of inspection words is small and the error detection capability is not sufficient, this becomes a problem. In the case of FIG. 1, the number of inspection words is 2
Therefore, if there are three or more errors, missed detection may occur. In order to prevent the erroneous correction due to such a missed detection, the error correction of the above (1) may be performed even when N (F ₀ ) = 4. In this case, it is possible to detect even an error to which no flag is added, and it is possible to correct if there is one or less error to which a flag is added. However, in the error correction of the above (1), FIG.
When there are three or more errors as described above, the correction becomes impossible and the correction capability deteriorates. An object of the present invention is to provide an error correction method and apparatus capable of maximally utilizing the error correction capability of a double-coded error correction code. In order to achieve the above object, the present invention provides a second coded codeword which constitutes a double coded codeword.
Error detection / correction is performed by the first decoding using the second check word, and at the same time, a flag is added to the second code block having an error. Then, for the first code block forming the codeword after this first decoding has been done and deinterleaved,
When error detection is performed by the second decoding using the first check word, and the number of error words detected in the first code block is equal to or less than the number determined by the minimum distance d _{1 of the first} check word. Corrects the detected error word,
If this number is exceeded, 2S ₂ + E ₂ ≤d ₁ -1 (where S ₂ is the first error code) depending on the number E _{2 of} error words to which the flag is added in the first code block. 1 of the combination of S ₂ and E ₂ which is determined by the number of words to which no flag is added among the error words detected in the code block
Select one and detect the error word in which the flag of S ₂ or less is not added in this selected combination, and S ₂ +
Flag of up to _two E performs correction of error words that are added with the error words that are not added. In the second decoding using the first check word subsequent to the first decoding, if the number of detected words cannot be corrected by the first check word, the first decoding is performed. The optimum decoding method is selected according to the number of words having the added flag, and the decoding is performed again by this selected decoding method. As a result, the error correction capability and the error detection capability are improved, and excellent error correction is performed. The following is a description of the case where the embodiment of the present invention is applied to the error correction code shown in FIG. As described above, as the decoding of the error correction code shown in FIG. 1, C ₁ decoding can correct 1 symbol error, and C ₂ decoding can correct 2 symbol error and 1 symbol error. It is possible to perform three types of decoding, that is, correction of erasure of symbols and correction of erasure of 4 symbols. Therefore, for C ₁ decoding, error detection and 1-symbol correction are performed as in FIG. 2, and an F ₀ flag is added when an error is detected and an F ₁ flag is added when an error cannot be corrected. To do. Next, the procedure of C ₂ decoding will be described with reference to FIG. (1) First, decoding is performed with S = 2 and E = 0. As a result, it is possible to detect error positions up to two symbols at arbitrary positions. When it is determined that the number of errors is two or less, it is determined based on the presence or absence of a flag added to the error position and the number of flags in the block, and if there is little possibility of erroneous correction, correction is performed. If there is a possibility of erroneous correction, it cannot be corrected. (2) In the case where it is determined that there are three or more errors in (1) and the number of cut F ₀ flags is 3, 2 symbols out of 3 symbols to which the F ₀ flag is added are erased, S =
Decoding of 1 and E = 2 is performed. Then, when it is determined that only the symbol to which the flag is added has an error, the correction is performed using the result of this decoding. (3) When it is judged that there are three or more errors in (1) and the number of F ₀ flags is 4, it is assumed that 4 symbols to which the F ₀ flag is added are lost and S = 0, E = 4 is decoded and 4 symbols are corrected. (4) When it is determined that there are three or more errors in (1) and the number of F ₀ flags is 5 or more, if the number of F ₁ flags is 3, the F ₁ flag is added. Decoding of S = 1 and E = 2 is performed by erasing 2 out of 3 symbols. When it is determined that only the symbol to which the F ₁ flag is added has an error, the result of this decoding is used for correction. (5) When it is judged that there are three or more errors in (1) and the correction conditions of (2) to (4) are not satisfied,
Uncorrectable. As described above, by performing the decoding twice using the different decoding methods, it is possible to correct even if the decoding by one circuit is uncorrectable or erroneous. For example, even in the cases of FIGS. 4 and 5 described above, FIG. 4 can be corrected by (3) and FIG. 5 can be corrected by (1). (1)
If the error cannot be corrected by (5) to (5), as shown in FIG. 6, when N (F ₀ ) = 4, there are two or more errors to which a flag is added and there is an error to which a flag is not added. Only. Further, even in such a case, if the number of decoding times is increased, it is possible to correct up to three arbitrary errors by performing S = 1 and E = 2 decoding at different erasure positions. . In addition, in (1) to (5), it was explained that the decoding is performed twice, but two different types of decoding are performed at the same time, and the optimum decoding method is determined after the decoding.
Error correction may be performed according to the decoding result of the decoding method determined to be optimum. FIG. 8 illustrates the embodiment shown in FIG. 7 in more detail. In the figure, 2 (F ₀ ) and 2 (F ₁ ) are C
₂ Error position detected by decoding and F ₀ flag or F ₁
It is the number that matches the position where the flag is added. Further, F is an uncorrectable flag added to the symbol determined to be uncorrectable. When F = 1, the flag is added to all symbols in the C ₂ block, and when F = F ₀ ,
The F flag is added only to the symbols to which the F ₀ flag is added. The symbols to which the F flag is added are subjected to error correction by means of mean value interpolation or the like during reproduction. As shown in FIG. 8, by checking the decoding result with the F ₀ flag and the F ₁ flag added in the C ₁ decoding, the probability of error correction can be reduced, and both the error correction capability and the error detection capability can be reduced. Excellent error correction can be performed. The error correction method of the present invention is not limited to that shown in FIG.
Applicable to both. 9 and 10 show another embodiment of the error correction method according to the present invention, in which 11 is an information word, 12 is a first check word, and 13 is a second check word. The encoding first consists of 24 information words W.
The first check words Qi, 0 to Qi, 7 are added to i, _{0 to} Wi, ₂₃ (i = _{0 to} 29) to form a C ₂ block 15. Next, 30 information words W ₀ , j ...
W ₂₉ , j (j = 0 to 23) or 30 first check words Q ₀ , k to Q ₂₉ , k (k = 0 to 7) to the second check word P ₀ j, P ₁ j (j = 0 to 31) is added, and C
₁ block 14 At the time of decoding, after error detection and error correction are performed in the C ₁ block, error detection and error correction are performed in the C ₂ block. The C ₁ block uses a Reed-Solomon code having a code length of 32, the number of check words 2 and a minimum distance of 3. Therefore, error correction of 1 symbol is possible, and C ₁
Decoding is the same as in the case of FIG. The C ₂ block uses a Reed-Solomon code with a code length of 32, the number of check words of 8 and a minimum distance of 9. Therefore, error correction can be performed within the range of 2S + E ≦ 8. The procedure of C ₂ decoding will be described below with reference to the flow chart shown in FIG. (1) First, decoding is performed with S = 4 and E = 0. This makes it possible to detect an error position of 4 symbols at an arbitrary position. When it is determined that the number of errors is 3 or less, or when the number of errors is 4 and the possibility of error correction is low, error correction is performed. (2) When it is judged that there are 5 or more errors in (1) and the number of F ₀ flags is 4, it is assumed that 4 symbols with the F ₀ flag added are lost and S = 2, E = 4 is decoded, and if the number of detected errors is 2 or less, error correction is performed. [0041] (3) (1) is determined that an error is 5 or more in and when the number of F ₀ flag is five, S = 1 to 5 symbols F ₀ flag is added as a loss, E = 5 is decoded, and if the number of detected errors is 1 or less, error correction is performed. (4) When it is determined that there are 5 or more errors in (1) and the number of F ₀ flags is 6, the symbols to which the F ₀ flag is added are regarded as erasures and S = 1, E = 6. Is decoded, and if the number of detected errors is 1 or less, error correction is performed. [0043] (5) is determined that an error is 5 or more in (1), and when the number of F ₀ flag is seven, S 6 symbols of the seven symbols F ₀ flag is added as a loss
= 1 and E = 6 are decoded, and if it is determined that only the symbol to which the flag is added has an error, the error is corrected. The error in (6) (1) is determined to 5 or more, and F ₀ If the number of flags is eight, F ₀ S = 0 flag 8 symbols is added as a loss, E = 8 is decoded and 8 symbols are corrected. (7) If the correction conditions of (1) to (6) are not satisfied, the correction is impossible. Next, an embodiment of an error correction device for performing error correction by the error correction method of the present invention will be described with reference to FIG. FIG. 11 is a block diagram showing such an error correction device. 17 to 19 are bus lines, 20 is a syndrome generation circuit, 21 and 22 are ROMs, and 25, 27 and 2.
9 is a RAM, 24 is an arithmetic circuit, 26 is a counter, 28 is a comparison circuit, 30 is a condition judging circuit, 31 is a program RO.
M and 32 are address counters. This device is composed of three bus lines, circuits connected to the bus lines, and a control circuit for controlling the operation of each circuit by a program. In FIG. 11, a bus line 17 is a data bus for exchanging data such as received signals and error patterns, a bus line 18 is a location bus for exchanging data such as a data position, and a bus line 19 is data. It is a flag bus for exchanging data of flags to be added. A data input / output terminal 38, a location input / output terminal, and a flag input / output terminal 40 are connected to each bus. The syndrome generation circuit 20 generates a syndrome based on the received signal input from the data input / output terminal 38. The arithmetic circuit 24 is the syndrome generation circuit 2
The calculation for obtaining the error position and the error pattern is performed by the syndrome generated by 0. The arithmetic circuit 24 performs multiplication, division and addition on GF (2m). The RAM 25 includes a shindom and an arithmetic circuit 24.
It is for storing the calculation result in. Reference numeral 23 is an 8-input OR circuit for determining whether or not the data on the data bus 17 is "0". The ROMs 21 and 22 are R for converting data between the data bus 17 and the location bus 18.
OM. That is, on the data bus 17, data is handled in vector representation, and the location bus 18
It is treated as a definite expression. Therefore, when exchanging data between the data bus 17 and the location bus 18, it is necessary to convert the data by the ROM 21 or the ROM 22. The counter 26 counts the number of flags in one block. In the second decoding, the counter 2
6, the number of F ₀ and F ₁ is counted, and the number is counted by the comparison circuit 28.
The number of words to be corrected is determined by comparing with a predetermined number, or it is determined whether or not the correction is performed or the correction cannot be performed without the correction. The RAM 27 is for storing the number of flags counted by the counter 26 and the error position. Further, the comparison circuit 28 is used for comparing the number of flags described above with a predetermined number, and comparing data with a constant during the decoding process. The RAM 29 has a flag F ₀ , which indicates the result of the first decoding added to the data in the second decoding.
F ₁ is stored. The status of the flag stored in the RAM 29 is used to check the presence / absence of the flag at the error position obtained by the decoding. The condition judging circuit 30 judges whether or not to branch the program based on the result judged by the OR circuit 23 and the comparison circuit 28 and the state of the flag stored in the RAM 29. The program ROM 31 stores a program for controlling each circuit described above to perform decoding. Reference numeral 33 is a signal that determines the address of the RAM and the constants that are input to each bus line and the comparison circuit. Reference numeral 34 is a signal that determines a condition for branching a program.
Is compared with the conditions of the OR circuit 23, the comparison circuit 28, the RAM 29, etc. to decide whether or not to branch. 35
Is a signal that determines the branch destination when branching. Also, 3
Reference numeral 6 is a signal for controlling a buffer and a register connected to each bus. The counter 32 controls the program address. The counter 32 advances the address of the program ROM 31 by the clock input from the master clock input 41 and executes the program. When branching the program, the branch instruction 37 loads the branch destination address 35 into the counter to branch the program. The input terminal 42 is for inputting a signal for resetting the counter 32 when the program is started. As a procedure for error correction, first, a received signal is input, a syndrome is generated, the number of flags is counted in the C ₂ decoding, and the RAM 29 in the flag state is used.
To remember. Next, decoding is performed by a program, an error position and an error pattern are obtained, and error data is corrected. Further, when it becomes uncorrectable in C ₁ decoding and C ₂ decoding, a flag to be added to the data is output from the flag input / output 40. As described above, the error correction device of the present invention uses the system of controlling each circuit by a program, the circuit scale is small, and only the program is changed for different error correction codes. Can be dealt with. As described above, according to the present invention,
The capability of the double-coded error correction code can be utilized to the maximum extent, and the error correction capability and the error detection capability can be improved.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a diagram showing a data array of double-coded error correction codes. FIG. 2 is a flowchart showing C ₁ decoding. FIG. 3 is a flowchart showing conventional C ₂ decoding. FIG. 4 is a diagram showing an error pattern when there are four flags in a C ₂ block. FIG. 5 is a diagram showing an error pattern when there are four flags in a C ₂ block. FIG. 6 is a diagram showing an error pattern when there are four flags in a C ₂ block. FIG. 7 is a flowchart showing C ₂ decoding according to the present invention. FIG. 8 is a diagram showing the flowchart of FIG. 7 in more detail. FIG. 9 is a diagram showing another example of a double-coded error correction code. 10 is a flowchart showing C ₂ decoding in the present invention when applied to the error correction code shown in FIG. 9. FIG. 11 is a block diagram showing an embodiment of an error correction circuit according to the present invention. [Description of Reference Signs] 20 syndrome generation circuit 21, 22 ROM 23 OR circuit 24 arithmetic circuit 25, 27, 29 RAM 26 counter 28 comparison circuit 30 condition determination circuit 31 program ROM 32 address counter

   ─────────────────────────────────────────────────── ─── Continued front page    (72) Inventor Keiji Noguchi             292 Yoshida-cho, Totsuka-ku, Yokohama-shi, Kanagawa             Ceremony Company Home Appliance Research Laboratory, Hitachi, Ltd. (72) Inventor Takao Arai             292 Yoshida-cho, Totsuka-ku, Yokohama-shi, Kanagawa             Ceremony Company Home Appliance Research Laboratory, Hitachi, Ltd.

Claims (1)

[Claims] 1. A large number of first code blocks each including a plurality of information words and a first check word whose minimum distance added to the plurality of information words is d ₁ (d ₁ is an integer of 2 or more) For the data block consisting of
Of a plurality of data blocks each consisting of a word selected from each of the code blocks and a second check word having a minimum distance d ₂ (d ₂ is an integer of 2 or more) added to the data sequence. Error detection / correction of the second code block by using the second check word in the error method for decoding the code word doubly encoded so that the second code block of In addition to performing the error detection / correction, a flag is generated for each second code block for which error detection is performed, and error detection is performed for the second code block using the second check word. Error detection is performed on the first code block of the corrected data block using the first check word, and the number of detected error words is 2S ₁ ≦ d ₁ −
When the predetermined number of words S _{1 is 1} or less, the error value of the word detected as an error using the first check word is obtained and corrected, and the word is detected using the first check word. If the number of error words generated exceeds S _{1, the} number of 2S ₂ + E ₂ ≦ d ₁ −1 is satisfied, depending on the number of words added with the flag included in the first code block. Select one of a plurality of combinations of the number of words E ₂ with a flag and the number of words S ₂ without a flag, and assume that the E ₂ words have an error. Again, the first check word is used to detect S ₂ or less errors other than the word having the error, and up to S ₂ + E ₂ words having the error and the errors are detected. The feature is that the error value of the word is obtained and corrected. Error correction method that. 2. A plurality of combinations of the number E _{2 of} words with the flag added and the number S _{2 of} words with no flag added include a case where S ₂ is 0. The error correction method according to item 1. 3. From each of the first code blocks, for a data block consisting of a number of first code blocks composed of a plurality of information words and a first check word added to the plurality of information words. A codeword that is doubly encoded so that a plurality of second code blocks are formed by a data series consisting of selected words and a second check word added to this data series. In the error correction device for decoding, the syndrome generation circuit that generates the syndrome from the information word and the check word, the flag generation device that adds a flag for each code block, and the information word and the check word in the code block. A counter that counts the number of flags that are present, and a memory that stores the positions of the flags that are added to the above information words and check words A circuit and an arithmetic circuit that performs an error correction operation using the syndrome generated by the syndrome generation circuit and the position information of the flag stored in the storage circuit, and corrects the error of the first code block. When you do
Error detection is performed using the first check word, and when the number of detected error words is equal to or less than a predetermined word number S ₁ which is 2S ₁ ≦ d ₁ -1, the first check word is set to If an error value of a word detected as an error is obtained and corrected, and the number of error words detected using the first check word exceeds S ₁ , then the first code block is used. the number of words which the flag is added _{contained, 2S 2 + E 2 ≦ d} 1 -1 a is _two words the number E of the flag is added and the flag is not added the number of words S ₂ of Select one of the combinations,
If there is an error in the E ₂ words,
Error of not more than S ₂ other than the word having the above-mentioned error is detected by using the check word of the above, and the error value of the word having the above-mentioned error up to S ₂ + E ₂ and the word having the error detected An error correction device characterized in that the correction is performed by obtaining